Désambiguïsation – Disambiguation
In darkness with a total silence, on a big screen in black and white, words appear and float in a tridimensional space.
Words leave a sentence to catch up with another one, then appear again. They move all the time and change as time passes by.
These words seems strangely current and are organized in sentences devoid of meaning, but grammatically and syntactically accurate.
The installation Désambiguïsation functions in real time. An algorithm continuously gathers texts from the information streams of different websites on the Internet. Before presenting a word on the screen, it identifies its grammatical structure in order to determine its place in the sentence. The process of semantic disambiguation then allows words to be organized together into a sentence by reducing the plurality of their meanings. The installation thus automatically generates a text in natural language.
This process of computational linguistics is used in human-computer communication for voice recognition, automatic translation, and spelling corrections, in particular. This work showcases language at work. Because of their particular behavior, the words, despite being managed by machines, seem to seek out their place, to hesitate.
The spectator as observer has no knowledge of this process and cannot anticipate the development of events. Just like the computer, we will also naturally process the text, trying to give meaning to what we read, to the rearrangements we observe. We enter into a relationship with the computer.
Jeff Guess voluntarily chose to sample the text sources from news websites since the language used is rather telegraphic. In the work, these words take on a whole new turn of phrase, and through the game of polysemy, poetry is born of the words.
Jeff Guess (1948-), a photographer and movie director, first worked on the ties between fixed images and images in movement during performances and with the help of experimental research on new technologies. His research progressively led him to a work on algorithms to describe this link, which is centered on the question of language, whether it is through image or text.